Различия

Показаны различия между двумя версиями страницы.

--- doc:os:freebsd:samba:prepare_zfs_pool_for_samba [d.m.Y H:i] – dbehterev
+++ doc:os:freebsd:samba:prepare_zfs_pool_for_samba [d.m.Y H:i] (текущий) – dbehterev
@@ Строка 76: / Строка 76: @@
 </code>
-TAG: {{tag> samba FreeBSD}}
+====== Выдержки по тюнингу ZFS ======
+В сети натолкнулся на статью, к сожалению не помню ссылку на автора:
+Система
+=======================
+  * Case - Supermicro SC733T-645B
+  *   MB - Supermicro X7SBA
+  *  CPU - Intel Core 2 Duo E8400
+  *  RAM - CT2KIT25672AA800, 4GB ECC
+  *  RAM - CT2KIT25672AA80E, 4GB ECC
+  * Disk - Intel X25-V SSD (ada0, boot)
+  * Disk - WD1002FAEX (ada1, ZFS "data" pool)
+  * Disk - WD2001FASS (ada2, ZFS "backups" pool)
+Samba
+=======================
+В smb.conf:
+  [global]
+  socket options = TCP_NODELAY SO_SNDBUF=131072 SO_RCVBUF=131072
+  use sendfile = no
+  min receivefile size = 16384
+  aio read size = 16384
+  aio write size = 16384
+  aio write behind = yes
+ZFS пулы
+=======================
+  pool: backups
+ state: ONLINE
+ scrub: none requested
+config:
+        NAME        STATE     READ WRITE CKSUM
+        backups     ONLINE       0     0     0
+          ada2      ONLINE       0     0     0
+errors: No known data errors
+  pool: data
+ state: ONLINE
+ scrub: none requested
+config:
+        NAME        STATE     READ WRITE CKSUM
+        data        ONLINE       0     0     0
+          ada1      ONLINE       0     0     0
+errors: No known data errors
+Тюнинг ZFS
+=======================
+Your tunings here are "wild" (meaning all over the place).  Your use
+of vfs.zfs.txg.synctime="1" is probably hurting you quite badly, in
+addition to your choice to enable prefetching (every ZFS FreeBSD system
+I've used has benefit tremendously from having prefetching disabled,
+even on systems with 8GB RAM and more).  You do not need to specify
+vm.kmem_size_max, so please remove that.  Keeping vm.kmem_size is fine.
+Also get rid of your vdev tunings, I'm not sure why you have those.
+My relevant /boot/loader.conf tunings for 8.2-RELEASE (note to readers:
+the version of FreeBSD you're running, and build date, matters greatly
+here so do not just blindly apply these without thinking first):
+  # We use Samba built with AIO support; we need this module!
+  aio_load="yes"
+  # Increase vm.kmem_size to allow for ZFS ARC to utilise more memory.
+  vm.kmem_size="8192M"
+  vfs.zfs.arc_max="6144M"
+  # Disable ZFS prefetching
+  # http://southbrain.com/south/2008/04/the-nightmare-comes-slowly-zfs.html
+  # Increases overall speed of ZFS, but when disk flushing/writes occur,
+  # system is less responsive (due to extreme disk I/O).
+  # NOTE: Systems with 8GB of RAM or more have prefetch enabled by
+  # default.
+  vfs.zfs.prefetch_disable="1"
+  # Decrease ZFS txg timeout value from 30 (default) to 5 seconds.  This
+  # should increase throughput and decrease the "bursty" stalls that
+  # happen during immense I/O with ZFS.
+  # http://lists.freebsd.org/pipermail/freebsd-fs/2009-December/007343.html
+  # http://lists.freebsd.org/pipermail/freebsd-fs/2009-December/007355.html
+  vfs.zfs.txg.timeout="5"
+sysctl tunings
+=======================
+Please note that the below kern.maxvnodes tuning is based on my system
+usage, and yours may vary, so you can remove or comment out this option
+if you wish.  The same goes for vfs.ufs.dirhash_maxmem.  As for
+vfs.zfs.txg.write_limit_override, I strongly suggest you keep this
+commented out for starters; it effectively "rate limits" ZFS I/O, and
+this smooths out overall performance (otherwise I was seeing what
+appeared to be incredible network transfer speed, then the system would
+churn hard for quite some time on physical I/O, then fast network speed,
+physical I/O, etc... very "bursty", which I didn't want).
+  # Increase send/receive buffer maximums from 256KB to 16MB.
+  # FreeBSD 7.x and later will auto-tune the size, but only up to the max.
+  net.inet.tcp.sendbuf_max=16777216
+  net.inet.tcp.recvbuf_max=16777216
+  # Double send/receive TCP datagram memory allocation.  This defines the
+  # amount of memory taken up by default *per socket*.
+  net.inet.tcp.sendspace=65536
+  net.inet.tcp.recvspace=131072
+  # dirhash_maxmem defaults to 2097152 (2048KB).  dirhash_mem has reached
+  # this limit a few times, so we should increase dirhash_maxmem to
+  # something like 16MB (16384*1024).
+  vfs.ufs.dirhash_maxmem=16777216
+  #
+  # ZFS tuning parameters
+  # NOTE: Be sure to see /boot/loader.conf for additional tunings
+  #
+  # Increase number of vnodes; we've seen vfs.numvnodes reach 115,000
+  # at times.  Default max is a little over 200,000.  Playing it safe...
+  kern.maxvnodes=250000
+  # Set TXG write limit to a lower threshold.  This helps "level out"
+  # the throughput rate (see "zpool iostat").  A value of 256MB works well
+  # for systems with 4GB of RAM, while 1GB works well for us w/ 8GB on
+  # disks which have 64MB cache.
+  vfs.zfs.txg.write_limit_override=1073741824
+TAG: {{tag> samba FreeBSD ZFS}}